Professional Documents
Culture Documents
1.determining The Impact of Distributed Generation On Power Systems
1.determining The Impact of Distributed Generation On Power Systems
Introd uction
Distributed Generation (DG) is predicted to play an increasing role in the electric power system of the near
fiture. Distributed generation is by definition that which is of limited size (roughly 10 MW or less) and
interconnected at the substation, distribution feeder or customer load levels. DG technologies include
photovoltaics, wind turbines, fuel cells, small 8z micro sized turbine packages, stirling-engine based
generators, and intemal combustion engine-generators. These technologies are entering a period of rapid
expansion and commercialization. For example, commercial-grade residential-scale fuel cells will arrive
within the next 2 or 3 years with expectations of millions of installed units within the next 10 years.
Photovoltaic programs are aggressively growing with plans to have nearly one million rooftops equipped
with PV within a decade. Wind farms are springing up worldwide with more than 3000 MW of new wind
capacity to be installed in the year 2000 alone. Even conventional internal combustion engine-generator
and small gas turbine installations are increasing as commercialhdustrial electricity users install them for
reliability and peak shaving applications. In fact, studies have predicted that distributed generation may
account for up to 20% of all new generation going online by the year 2010. With so much new distributed
generation being installed, it is critical that the power system impacts be assessed accurately so that these
DG units can be applied in a manner that avoids causing degradation of power quality, reliability, and
control of the utility system. This paper is the first in a two part series reviewing some of the concerns of
DG application and methodologies for assessing the system impacts. This first paper focuses primarily on
radial distribution systems while the second paper focuses on low voltage distribution networks and
dynamic stability issues.
Analyzing DG Impacts
DG on Distribution Systems
Distribution systems, whether they are radial type systems found in rural or suburban areas, or network
type systems found in urban areas are generally designed to operate without any generation on the
distribution system or at customer loads. The introduction of generation sources on the distribution system
can significantly impact the flow of power and voltage conditions at customers and utility equipment.
These impacts may manifest themselves either positively or negatively depending on the distribution
system operating characteristics and the DG characteristics. Positive impacts are generally called “system
support benefits,” and include:
Achieving the above benefits is in practice much more difficult than is often realized. The DG sources must
be reliable, dispatchable, of the proper size and at the proper locations. They must also meet various other
operating criteria. Since many DGs will not be utility owned or will be variable energy sources such as
As an example, if a DG unit is applied just downstream of a voltage regulator or LTC transformer that is
using considerable line drop compensation (as is shown in Figure I), then the regulation controls will be
unable to properly measure feeder demand. Figure 1 shows the voltage profiles that can occur under these
conditions with and without DG. Note that with distributed generation the voltage becomes lower on the
feeder. In this example the voltage is reduced because the DG reduces the observed load at the line drop
compensator control. This confuses the regulator into setting a voltage lower than is required to maintain
adequate service levels at the tail end of the feeder. This is the opposite effect of “voltage support”, a
commonly touted benefit of DG. Movement of the DG unit to the upstream side of the regulator (if
possible) or additional regulator controls to compensate for the DG output are solutions to this problem, but
they bring their own set of issues.
I
~~~~
Substation
r---------------
i
I
Gen
- Area of Low
Voltago
p:
i
I
Feeder voltage profile at peak
demand with DG Unit Off
-
------z-,- - - .--._-_.*.................................................
...............................................................................................................................
L v v r A N X l Limit Row. A Linil(99.i ~ ~ 1 1 1 1 e I -----____
Substation End of
Bur Feeder
Location
Figure 1 - Example of DG unit interfering with voltage regulation1 on a distribution feeder. Line drop
compensation must be employed a t the LTC control to result in thie indicated voltage profiles.
The preceding examples have shown ways that both high or low service voltage can occur due to the
incompatibility of DG with the radial power flow based voltage regulation approach used on most utility
systems. As a result, the DG influence on voltage for any potential application shodd always be assessed to
insure that no customers will be adversely impacted.
An accurate way to analyze the voltage regulation behavior of a system with DG is to run a power flow
simulation using software capable of analyzing multiple sources on the distribution system and the
operation of voltage regulators. The phase angle of the injected DG current relative to the utility system
voltage must be considered as well as the impedance characteristics of the line. In this analysis, it is
important to recognize that the DG injected power may result in a voltage that is within limits at the DG
site but could be out of limits farther downstream.
To determine if the DG will cause a significant impact on the feeder voltage, the size and location of the
DG, the voltage regulator settings, and impedance characteristics of the line must be considered. If line
drop compensation is used by the regulator, then DG units interconnected within the regulator’s zone and
downstream of the constant voltage point (CVP) will support (increase) the feeder voltage below the CVP.
Those above the CVP will lower the voltage below the CVP. If voltage support is the key reason for using
DG, then DG placement downstream of the CVP is crucial to meeting this objective (the farther
downstream of the CVP, then the greater the support). The values of line-drop-compensation utilized will
determine the CVP location on the feeder. Note that when no line-drop-compensation is used, the CVP is at
the regulator device itself.
In essentially all cases the impact on the feeder primary will be negligible for any individual residential
scale DG unit (<lOkW). However, when the aggregate capacity of many small units deployed reaches a
critical threshold or when the capacity of a single unit is large enough, then voltage regulation studies are
desirable to insure that the feeder voltage will be maintained within appropriate limits. The aggregate DG
capacity threshold for which studies become appropriate depends on many factors. However, a reasonable
rule-of-thumb is that if the injected current (measure at the primary level) is less than 5% of the feeder
loading at the interconnection point and if all customers had satisfactory voltage prior to the addition of the
DG, then a voltage problem on the primary is unlikely. Much above 5% injection and there is more reason
to worry about potential impacts. For shared secondaries with DG even a small generator that injects less
than 5% at the primary level could pose a voltage regulation risk to customers sharing the secondary. Thus,
analysis of voltage conditions will almost always need to consider impacts on the secondary where the DG
is located.
Distributed generation will also impact losses on the feeder. DG units can be placed at optimal locations
where they provide the best reduction in feeder losses. Siting of DG units to minimize losses is like siting
capacitor banks for loss reduction. The only difference is that the DG units will impact both the real and
reactive powerflow. Capacitors only impact the reactive power flow. Most generators will be operated
between 0.85 Lagging and 1.0 power factor, but some inverter technologies can provide reactive
compensation (leading current). A good loadflow analysis software should be able to model the effects on
system losses. On feeders where losses are high, a small amount of strategically placed DG with an output
of just IO-20% of the feeder demand can have a significant loss reduction benefit for the system.
Unfortunately, most utilities don’t have control over the siting locations, since DG is usually customer
Larger DG units must be sited with consideration of feeder capacity limits. In some cases overhead lines
and cables may be thermally limited meaning that the DG can inject power that exceeds the line’s thermal
limit without causing a voltage problem on the feeder. The power flow analysis should%“flag”the locations
where capacity constraints will be an issue from a thermal as well as a voltage perspective. In general, a
DG at a location that is thermally limited is not connected at the optimal point from a “power loss”
perspective.
Voltage Flicker
DG may cause noticeable voltage flicker. Flicker can be either a simpk: issue or a complex issue as far as
its analysis and mitigation are concerned. From the simple perspective. it can be the result of starting a
machine (e.g. induction generator) or step changes in DG output which result in a significant voltage
change on the feeder. If a generator starts, or its output fluctuates fiequently enough, flicker of lighting
loads may be noticeable to customers. One approach is to determine the magnitude and number of changes
of voltage occurring per unit time and see if these are above the visibility or irritation threshold levels of the
GE flicker curve[*](see Figure 2). If above the thre:shold, or if customer complaints occur, mitigation must
be considered. Mitigation approaches include reduced voltage starts on1 induction generators as well as
speed matching. Synchronous generators might require tighter synchronization and voltage matching.
Inverters might be controlled to limit inrush currents and changes in output levels. A less technical
approach to reduce flicker involves placing constraints on when and how often DG operators may start and
change the output of DG systems. In the case of wind and solar energy systems, the outputs will fluctuate
significantly as the sun and wind intensity change. Fortunately, these fluctuations tend to be smoother than
step changes associated with the GE flicker curve. For this reason, the GE flicker curve is a conservative
curve for assessing some types of solar and wind induced voltage fluctuations.
Figure 2 - GE Flicker Curve showing the “borderline of irritation” (see IEEE 519-1992 for
further details)
Determination of the risk of flicker problems due to basic generator starting conditions or output
fluctuations is fairly straightforward using the flicker curve approach, Imticularly if the rate of these
fluctuations is well defined, the fluctuations are “step” changes and there are no complex dynamic
interactions of equipment. The dynamic behavior of machines and their interactions with upstream voltage
regulators and generators can complicate matters considerably. For ex:unple, it is possible for output
fluctuations of a DG (even smoother ones @omsolar or wind systems) to cause hunting of an upstream
regulator and, while the DG fluctuations alone may not create visible flicker, the hunting regulator may
Harmonics
Distributed generators may introduce harmonics. The type and severity will depend on the power converter
technology and interconnection configuration. In the case of inverters, there has been particular concem
over the possible harmonic current contributions they may make to the utility system. Fortunately, these
concems are in part due to older SCR type power inverters that are line commutated and produce high
levels of harmonic current. Most new inverter designs are based on IGBTs that use pulse width modulation
to generate the injected "sine" wave. These newer inverters are capable of generating a very clean output
and they should normally satisfy the IEEE 519-1992 requirements (see Table 1 below).
A l l o w e d L e v e l R e l a t i v e to F u n d a m e n t a l
Harmonic Order ( 0 d d h a r m o n i c r )*
<11'b 4%
Table 1. Harmonic Current Injection Requirements for Distributed Generators Per IEEE
519-1992.
Rotating generators such as synchronous generators can be another source of harmonics. Depending on the
design of the generator windings (pitch of the coils), core non-linearity's, grounding and other factors, there
can be significant harmonics. Triple harmonics are additive in the neutral; and the third harmonic is often
the most prevalent. Synchronous generators are often specified with a 2/3 pitch for the windings since this
produces much less third harmonic than those with other pitches. Unfortunately a 213 pitch machine has a
lower impedance to third harmonic and may cause more harmonic current to flow from other sources
connected in parallel with it. The grounding arrangement of the generator and step-up transformer will play
a major role in limiting the feeder penetration of harmonics. Arrangements can be selected to block or
reduce third harmonic injection to the utility system. This would tend to confine it to the DG site only.
In general, harmonic contributions from DG units are probably not as much of an issue as some of the other
problems discussed in this paper. Nonetheless, there are specific instances where problems have been
identified such as levels of distortion which exceed the IEEE-5 19 standard, resonance with capacitor banks,
or problems with equipment that are sensitive to harmonics. In extreme cases, equipment at the DG site
may need to be derated due to added heating caused by harmonics. Any DG installation design should be
reviewed to determine its compliance with IEEE 5 19 and whether harmonics will be confined to the DG
site or also injected into the utility system. For larger DG units or cases involving complex harmonic
su
Breaker
,& Fault
Figure 3 - Fault contributions due to DG units 1,2 and 3 may increase the short circuit levels to the
point where fuse-breaker coordination is no longer achieved!
Typical short circuit levels of DG power converters are characterized in Table 2 and references [5][6][7]
and [SI.For inverters, the fault contributions will dlepend on the maxirnum current level and duration for
which the inverter manufacturer’s current limiter is set to respond. On some inverters fault contributions
may last for less than a cycle, in other cases it can be much longer. For synchronous generators, the current
contribution depends on the prefault voltage, subtransient and transient reactances of the machine, and
exciter characteristics. Induction generators can also contribute to faults as long as they remain excited by
any residual voltage on the feeder. For most induction generators, the significant current would only last a
few cycles and would be determined by dividing the prefault voltage by the transient reactance of the
machine, Even though a few cycles is a short time, it is long enough to impact fuse-breaker coordination
and breaker duties in some cases.
00-400% (duratior
~ Generator
Induction Generator or Selfbcited
S nchronous Generator
Starting at 500-100
- anddecayingto200-400%~cles
500-1 000% forJrst,kw cycles and decaying to
a negligible amount within 10
I
When a single generator is added to the system, a manual calculation of the peak fault currents based on
manufacturers data can be performed to screen for a serious impact on the existing short circuit levels. For
multiple generation devices scattered throughout the system or large generators, the only accurate approach
is to perform a software based short circuit analysis which correctly models the short circuit behavior of the
generators. In many cases, the DG units won’t pose a threat to existing coordination; only a relatively few
cases may require changes in protection settings.
Most U.S. distribution systems are four-wire-multigrounded-neutralsystems that are defined as being
effectively grounded with respect to the substation source. For a single line to ground fault, this
arrangement limits the voltage rise on unfaulted phases to about 125 to 135% of the prefault conditionr31141.
By IEEE definition, effective grounding means that the positive sequence reactance is greater than the zero
sequence resistance (XI>%) and the zero sequence reactance is less than three times the positive sequence
reactance (3X1>X,). Use of a DG source that does not appear as an effectively grounded source connected
to such systems may lead to overvoltages during line to ground faults on the utility system. This condition
is especially dangerous if a generation island develops and continues to serve a group of customers on a
faulted distribution system. Customers on the unfaulted phases could in the worst case see their voltage
increase to 173% of the prefault voltage level for an indefinite period. At this high level, utility and
customer equipment would almost certainly be damaged. Saturation of distribution transformers will help
slightly to limit this voltage rise. Nonetheless, the voltage can still become quite high (1 50% or higher). For
example, marginally rated lightning arresters would probably fail in less than 5 seconds if the voltage
reached just 150% of the prefault condition 14].
To avoid problems, all DG sources on multigrounded neutral systems that are large enough to sustain an
island should present themselves to the utility system as an effectively grounded source. If they do not, they
should use appropriate protective relaying to detect primary side ground fault overvoltages and quickly trip
off-line (instantaneous trip). The former approach is preferred since it limits by design the voltage swells
that the system will see during a fault. The latter approach, while used successfully in many installations,
could subject the customer to many cycles of severe overvoltage prior to the DG unit being cleared f?om
the system. And if the DG is not cleared quickly enough, equipment could be damaged.
Some common transformer configurations found at DG sites are shown Table 3. Not all of these provide
effective grounding. Use of an adjacent grounding transformer bank is one approach for creating a DG
source that behaves as an effectively grounded source when the step-up transformer does not provide
effective grounding. It should be noted that any additional grounding sources on the system, whether it is
by a grounding bank or by using an appropriate step-up transformer winding arrangement, will have the
effect of reducing (diverting) the zero sequence return current seen at the substation that arises from line to
ground faults on the feeder. The amount of current diverted is determined by the ratio of the DG grounding
source impedance to the substation grounding impedance. This effect has the impact of desensitizing the
Proper review of the generator characteristics and utility system design prior to installation of the DG unit
can insure that grounding compatibility problems are dealt with appropriately. Grounding compatibility is
crucial to insure that power quality and reliability are not degraded b-y the addition of DG units.
-
or I suitably sized primaryside grounding bank is provided I
Delta/Delta
Wye-groundeWye grounded
grounding bank is provided or if the generator is wye-
grounded and meets neutral impedance requirements
Wye-grounded/Delta Typically provides effective grounding regardless of
I generator grounding arrangement
Islanding
Islanding occurs when the distributed generator (or group of distributed generators) continues to energize a
portion of the utility system that has been separated from the main utility system (see Figure 4). This
separation could be due to operation of an upstream breaker, fuse, or automatic sectionalizing switch.
Manual switching or “open” upstream conductors could also lead to islanding. Islanding can occur only if
the generator(s) can self excite and sustain the loaid in the islanded section. In most cases it is not desirable
for a DG to island with any part of the utility systfembecause this can lead to safety and power quality
problems that will affect the utility system and lOi5dS. For example, if an island develops on a feeder during
standard reclosing operations, the islanded DG units will quickly drift out of phase with the utility system
during the “dead period.” Then, when a reclose occurs, the utility will connect out of phase with the island
if reclose blocking into an energized circuit is not provided at the breaker control. This can cause damage to
utility equipment, the DG units supporting the island, and customer 1,oads.Islanding also increases the
likelihood that DG sources may be allowed to subject the island to out of range voltage and frequency
conditions during its existence. And it can pose a serious safety threat during downed conductors and utility
repair operations since the public and utility workers may be exposed to circuits that otherwise would be
de-energized. Finally, islanding can hinder service restoration by requiring line crews to spend extra time
disabling the island conditions. This will impact reliability indices such as SAIDI.
This area of
primary stili
en erg ized
House with
Fuel Cell and
grid connected
inverter
-
Figure 4 Islanding can happen in a variety of circumstances. Proper protection must be provided to
avoid islanding in all typical applications of DG. Most utilities and the IEEE have standards
specifying the protection required to help prevent island development. Active anti-islanding is a
robust new protection approach being applied to inverters.
To prevent islanding, a DG unit operating in parallel with the utility system should in a timely manner
sense a significant voltage sag or discontinuity of service on the utility side and disconnect from the
system. Many utilities have standards that require that this time be about 10 cycles or less for serious feeder
disturbances (deep voltage sags or interruptions). IEEE has also developed standards; for example, IEEE
929 recommends that small photovoltaic inverters should initiate a disconnect within 6 cycles for such
events [’I. This is fast enough to allow the DG’s to disconnect prior to the fastest reclosing times used in the
industry.
Voltage and frequency relays are used as a means of anti-island protection. In most cases, if a generator
becomes islanded, it will not be able satisfy the sudden change in its load without a significant change in
voltage and/or frequency and the relays will trip the unit off line. This type of anti-islanding protection is
called “passive.” Passive protection can be fooled if the generator is able to carry the load of the island
without a substantial change in voltage or frequency so, as a further safeguard, many smaller inverters
today also use what is called “active” anti-islanding protection. One common “active” approach is for the
inverter to be “tuned” to operate while islanded at a frequency other than 60 Hz.While the utility signal is
present the inverter is “forced” to operate at 60 Hz.During an intemption, it will drift to its natural “tuned”
frequency once the “forcing utility voltage function” disappears (that is when it becomes islanded). In the
process of drifting to the natural inverter frequency, the unit will trip its internal frequency relays, which
are set to trip for a deviation from power frequency outside the range of 59.3 to 60.5 Hz. Active anti-
islanding is more robust than passive, but even it cannot guarantee that an island won’t develop in some
rare cases. Some active anti-islanding techniques developed by Sandia Laboratories are discussed in
reference [9].
Since islanding can cause severe voltage quality and reliability problems, the proper use and setting of anti-
islanding controls is one of the more important issues for DG installations. For small photovoltaic inverters
IEEE standards are already available which describe the settings and types of controls required [’I. For other
types of inverters and larger installations, the settings and controls required are still being defined. IEEE
has a new standard under development (IEEE P1547) which will address all DG sizes and technologies.
Most utilities and many state energy commissions also have standards that address the islanding issues.
Area
slupported
during utility
1 outage
Substation
Conclusions
DG has much potential to improve distribution system performance and it should be encouraged. However,
distribution system designs and operating practices are normally based on radial power flows and this
creates a special challenge to the successful introduction of distributed generation. This paper has described
a few of the issues that must be considered to insure that DG will not degrade distribution system power
quality, safety or reliability. This paper focused on radial systems, ahhough some of the issues discussed
are common to low voltage distribution networks. For a review of low voltage network considerations and
other issues related to DG dynamics, the second paper (Part 2) of this series is referenced [lo].
It is clear that detailed analysis and special equipment to prevent protilems will be required in many cases
where larger DG units are applied. However, it is important to recognize that due to the economics of small
DG installations, a detailed study can not be performed for every small DG installation. That would be too
much of a burden from a cost perspective on both the utility and DG operators (or whoever ends up paying
for the studies). To resolve this issue, DG applications need to be screened with a “very low cost
1. Those that can clearly be installed per codes and standards without any special requirements
2. Those that are borderline (may need some limited special studies or have limited additional
equipment requirements)
3. Those that are clearly in need of detailed system impact studies and addition equipment
requirements
The capacity rating alone of a DG unit is not sufficient information to determine whether the installation is
in Category 1,2 or 3. The factors that determine system impacts include not only the size, but the dynamic
characteristics of the generator, the characteristics of the distribution system at the point of interconnection
and the presence of other customers on the same secondary. As an example, it is possible for a 1000 kW
DG connected at a stiff 12 MVA substation bus to cause less impact on system voltage than a 25 kW DG
operating on the secondary of a 50 kVA distribution transformer serving many customers!
To avoid requiring costly impact studies for all but those applications that actually need them, proposed DG
projects can be screened based on factors such as:
These and other factors will determine the likelihood of causing a significant impact on the system and
adjacent customers. The screening can be done manually or it can be done automatically by software based
screening tools. A typical software screening tool would lead the user through a series of questions about
the utility system characteristics and the DG unit characteristics and would perform various calculations to
identify if the proposed DG installation needs a detailed study or can be installed without significant
concern of adverse impacts. Generally, the larger installations will most commonly need special studies
and attention but there will be smaller units that occasionally may require specialized study. Effective
screening tools of this nature that consider factors on both the utility side and DG side of the system are
under development and are strongly needed in the industry to streamline processing of an almost explosive
growth in DG interconnection applications and projects. This will help to promote DG and maintain
adequate safeguards to protect utility system reliability, safety and performance.
References
2. IEEE Recommended Practices and Requirements for Harmonic Control in Electric Power Systems,
IEEE Standard 5 19-1992, Published by Institute of Electrical and Electronic Engineers, 1992
3. IEEE Guide for the Application of Neutral Grounding in Power Systems, Parts 1-4, IEEE Std. C63.92
Parts 1-4, , Published by Institute of Electrical and Electronic Engineers, 1992
4. IEEE Guide for the Application of metal Oxide Surge Arresters for Alternating Current Systems, IEEE
Std. C62.22-1997, Published by Institute of Electrical and Electronic Engineers, 1998
5. Recommended Practice for Utility Interface of E'hotovoltaic (PV) S'ystems, IEEE Std. 929-2000,
Published by IEEE, 2000
6. IEEE Guide for Interfacing and Dispersed Storage and Generation Facilities with Electric Utility
Systems, ANSIAEEE Std. 1001- 1988, April 1989. Published by IEIEE.
7. Onan Liquid-Cooled Generator Set Application Manual (T-030), Onan Corporation, November 1993
8. Static Power Converters of 500 kW or Less Serving as the Relay Interface Package for Non-
Conventional Generators, IEEE 93THO 597-5 PWR, Published by IEEE, 1993.
9. G.A. Kern, R.H. Bonn, J. Ginn, S . Gonzalez, Results of Sandia Najional Laboratories Grid-Tied Inverter
Testing, Presented at the 2ndWorld Conference and Exhibition on I?hotovoltaic Solar Energy
Conversion, 6-10 July 1998, Vienna, Austria. (Available at Sandia S website www. sandia.gov/pv)
10. Philip Barker, Jim Feltes, David Smith, "Determining the Impact of Distributed Generation on Power
Systems: Part 2 - Low Voltage Network System.s and DG Dynamics